word alignment 0.00840488
alignment links 0.0062612499999999995
alignment link 0.00620405
word translation 0.00576708
alignment models 0.005760104
alignment loss 0.0055855489999999995
alignment decoding 0.0055396149999999995
alignment error 0.005490236
mbr alignment 0.0054348849999999995
alignment probabilities 0.005420082
several alignment 0.0053900779999999995
alignment quality 0.0053827879999999995
automatic alignment 0.005364918
alignment lattice 0.005358016
alignment errors 0.005317286
alignment process 0.0053154629999999994
bitext alignment 0.00531347
alignment experiments 0.005307065
reference alignment 0.005302127
alignment procedures 0.005282623
alignment perfor 0.005278459
generalized alignment 0.005262882
sure alignment 0.00525892
tence alignment 0.005256025
alignment alternatives 0.005250294
errorful alignment 0.005249402
erence alignment 0.005248961
alized alignment 0.005248258
alignment 0.00491679
word alignments 0.00481041
english word 0.004330431
word association 0.003942921
automatic word 0.003936218
null word 0.0038803920000000003
french word 0.003862725
matic word 0.0038487340000000004
likely word 0.0038398440000000002
word classes 0.003833333
word cluster 0.003829743
word fertility 0.00382685
word class 0.003825844
word clusters 0.0038254020000000003
unambiguous word 0.003821122
translation model 0.0034680099999999997
translation models 0.003122304
state translation 0.002937516
translation systems 0.00282989
machine translation 0.002823084
statistical translation 0.0027990479999999997
translation probabilities 0.0027822819999999996
translation quality 0.002744988
translation framework 0.002674934
translation 0.00227899
link set 0.0020846989999999998
link posterior 0.002048833
english words 0.001984531
english sentence 0.001930611
test data 0.0019225569999999999
mbr alignments 0.0018404150000000002
ment link 0.001778015
link sets 0.0017539209999999999
good alignments 0.001737906
consensus alignments 0.001727203
link weights 0.001710399
tion model 0.001698873
ous links 0.001689963
sure links 0.00168659
matic alignments 0.0016829640000000001
system performance 0.0016801799999999999
ence alignments 0.001669731
poor alignments 0.001658514
link weight 0.001647825
sentence pairs 0.0016145649999999999
test corpus 0.0016039119999999999
distortion model 0.00154783
simple model 0.001545215
fertility model 0.00152778
tility model 0.001525041
model architectures 0.001521983
french words 0.0015168249999999999
sentence pair 0.0015024889999999999
sentence length 0.001472854
french sentence 0.001462905
posterior probability 0.0014609599999999999
other loss 0.001390341
statistical models 0.0013633719999999998
baseline models 0.00135176
links 0.00134446
alignments 0.00132232
english side 0.001318313
linguistic features 0.001314801
english sen 0.0012683030000000001
posterior probabilities 0.001264865
modeling approach 0.001256615
initial state 0.001250816
decoder performance 0.0012491590000000001
finite set 0.0012469970000000001
error loss 0.0012422050000000001
various loss 0.00123186
automatic align 0.001230599
